Project-MONAI / Project-MONAI/model-zoo

Use CacheDataset for generative bundles.

Open
#544 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
338
Forks
106
PR merge metrics
No merged PRs in 30d

Description

Is your feature request related to a problem? Please describe.

For the diffusion bundle, monai.apps.DecathlonDataset is used by the generative bundles.

https://github.com/Project-MONAI/model-zoo/blob/f9334c14dbd779898b4ec04ac2f72554b486a668/models/brats_mri_generative_diffusion/configs/train_autoencoder.json#L143

https://github.com/Project-MONAI/model-zoo/blob/f9334c14dbd779898b4ec04ac2f72554b486a668/models/brats_mri_axial_slices_generative_diffusion/configs/train_diffusion.json#L105

It provides some convenience (auto-downloading datasets), but it makes overwriting the input datalist more difficult. Can we replace it with CacheDataset like the other bundles, which have more flexibility?

Describe the solution you'd like
Use CacheDataset for generative bundles.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the dataset definitions in models/brats_mri_generative_diffusion/configs/train_autoencoder.json and models/brats_mri_axial_slices_generative_diffusion/configs/train_diffusion.json, then compare them with the other bundles that use CacheDataset. Replace the generative bundle dataset configuration so the input datalist can be overridden, and confirm both bundle configs still load correctly.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
55/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.